We have some exciting updates from our work on “blending survival curves” for HTA — the paper is out in the print version of MDM, this week (though it’s come out as an e-print a little earlier in the year).

We have now a decent version of the companion package, which we aptly named blendR — Nathan has found a nice logo for the package, though I somehow wished we could use the gif instead…

The package can be installed using remotes and its GitHub repo and I’ve written a relatively short explainer here. In reality, it will be a fairly simple set of functions, because the blending process is basically just some algebra applied to two survival curves, to construct a synthesis that gives time-depending weights to the two components of the mixture…

But: we’re trying to make it as general as possible and with quite a few simplifying facilities for the user. For now, the basic inputs are in the form of fitted curves (to both the short-term data and some long-term process, which may be based on hard evidence or expert elicitation). But we allow these to be obtained through a range of useful packages (so all survHE’s flavours, but also flexsurv and potentially others too).

And the idea is to create all sorts of visualisation functions, like simple plots that automatically includes the blended curve, as well as depictions of the chosen weight functions and other bits and pieces which we are still working on — the package is fully functional, but we’re looking into further improvements… Vignettes and more examples to come shortly too!